PREVALENCE OF AND RISK-FACTOR PROFILE FOR PANCREATITIS IN MALE VETERANS WITH MIXED ALCOHOL AND DRUG ADDICTION

Abstract

Background and Aims: To determine the prevalence of and risk factor profile for pancreatitis among high-risk veterans. Methods: The study population consisted of 1409 black or white men with alcohol-related ICD-9 codes (291, 303, 305.00, 305.01, 305.02, 305.03) attending a large VA Outpatient Detoxification Program (1/2002-12/2003) who had systematically collected risk-factor information on alcohol use, drug use and smoking. Pancreatitis (at anytime through 6/2005) among study subjects was identified using ICD-9 codes (acute pancreatitis [AP] - 577.0, chronic pancreatitis [CP] - 577.1). Clinical records of all subjects with pancreatitis codes (n = 87) and 214 controls within this population (no ICD-9 codes for pancreatitis) were reviewed in a blinded fashion to verify pancreatitis diagnosis. Definite AP was diagnosed by presence of typical abdominal pain with amylase ± lipase elevation ≥3 times normal or imaging evidence of pancreatitis, and alcoholic CP based on the International workshop on CP criteria for definite diagnosis (Pancreas 1997;14:215-221). Results: Most subjects were middle aged (median 48 yrs, IQR 44-53) and 43.5% were whites. History of smoking (90%), current or past drug use (90%) and IV-drug use (22.3%) was common. Although 6.2% subjects had pancreatitis codes (n = 87, AP-50, CP-15, both-22), chart review verified only 3% cases (n = 42, AP-29, CP-5, both-8). After excluding 3 patients who also had gallstones on imaging, definite alcoholic pancreatitis was present in 39 of these 42 patients. The prevalence of pancreatitis increased to 4.2% if milder criteria to diagnose AP (typical abdominal pain with any enzyme elevation and negative imaging study) and CP (clinical picture consistent with CP but no morphological changes of CP as included in the definite criteria) were used. On bivariate analyses, compared to controls (no pancreatitis), pancreatitis subjects were significantly older (median age 53 vs. 48 yrs, p < 0.001), had more admissions to substance abuse program (median 4 vs. 2, p < 0.01), reported higher current alcohol use (median 17 vs. 12 drinks/d, p < 0.05) and lower drug dependence (p < 0.05). In multivariable logistic regression analyses, alcoholic pancreatitis associated positively with age (OR = 1.08, 95% CI = 1.04-1.12, p < 0.001) and number of substance abuse admissions (OR = 1.08, 95% CI = 0.995-1.18, p = 0.06). Conclusions: Pancreatitis prevalence in this high-risk population is at least 3%, but likely ranges between 3-6%. Many behavioral factors associate with pancreatitis. Our study provides preliminary data for future studies focusing on risk factor associations for pancreatitis in high-risk populations.